Saturday, August 4, 2012

IRS Submits Reformulated Treaty Request for Credit Suisse U.S. Clients (8/4/12)

An article in Tax Notes Today reports that the IRS has submitted a "new, reformulated" treaty request to the Swiss Federal Tax Administration under the Exchange of Information provision of the U.S./Swiss Double Tax Treaty.  Kristin A. Parillo and Marie Sapirie, IRS Submits New Treaty Request to Switzerland on Credit Suisse Data, 2012 TNT 151-6 (8/6/12).

The prior request was withdrawn after the Swiss Federal Administrative Court rejected it.  Per the article:
The court held that the request -- which didn't specifically identify the names of targeted bank clients but instead sought information on clients who had allegedly concealed U.S. income and assets with the assistance of bank employees -- was too vague, was based on search criteria indicating only tax evasion and not tax fraud, and was not sufficient to require cooperation by the Swiss tax authorities under the current Switzerland-U.S. tax treaty.
The TNT article cites a Swiss newspaper as indicating that (i) the reformulated request provides "a more detailed description of the manner in which the bank and clients concealed U.S. income and assets"; (ii) the "FTA directed Credit Suisse to inform affected clients by July 31 that their information is subject to disclosure under a treaty request made by U.S. authorities;" and (iii) "The new request affects fewer than 100 U.S. clients."

Other Articles:

U.S. Makes New Request to Swiss for Information on Alleged Tax Cheats ( 8/30/12), here.


  1. Jack,

    The TNT reference: "The new request affects fewer than 100 U.S. clients," seems to be inaccurate.

    Being the curious and exigent type, I found the original Swiss newspaper reference which states (translating as accurately I can) that the paper believes (as close a translation as possible to the words used) that 100 American clients had been notified they would be affected by the revised request. That is not the same as saying it "affects fewer than 100 U.S. clients." Obviously, many more could have been notified than the paper was aware of, which is highly likely to be the case I suspect.

    But my question is whether the DOJ would limit its request to a number rather than "specifically identify" the names in accordance with the ruling of the Swiss Court? Or is this just loose reporting where the DOJ has identified clients and the reporter is filling in the rest of what he perceives should be the story using his imagination rather than facts?

    Thanks for any enlightenment you can offer.


  2. Patrick,

    I don't think the IRS would, for example, make the request to send the top 100 Credit Suisse U.S. depositors / tax fraudsters. Rather, the IRS would identify characteristics that would identify tax fraud that the Swiss could / would then use to identify the specific Credit Suisse depositors.

    I seem to recall that Credit Suisse (and perhaps other Swiss banks) had turned over nonidentifying data about U.S. clients already. From that data, depending upon how deep and broad, the IRS could identify the characteristics that would relate to only a certain number of depositors. Thus, by using the Credit Suisse data set and identifying particular characteristics, the IRS and the Swiss / Credit Suisse would be able to determine the number involved. Now, I don't know that that is what happened, but if you give me a data set with a broad set of characteristics, I suspect that I could design a request and know roughly how many Credit Suisse depositors met the defined characteristics.

    Still, all this is guesswork.

    The key known point is that the Swiss now find it in their interests to respond to some such John Doe Treaty Requests.

    Thanks for all your comments,

    Jack Townsend

  3. I have gone to the NZZ article:
    and I believe it states that originally the US had asked for data on 650 account holders (and Credit Suisse has sent that data to Bern.) Of these, the revised request only covers 100 (I would assume all numbers are approximations) but however 150 clients of the original 650 already have had their information sent. So in a sense the horse is already out of the barn for them.

  4. I am trying to guess what the big-picture meaning is of far fewer client identities being revealed by CS compared to UBS. Maybe this means that CS did not engage in the type of outrageous behavior that UBS did (such as sending bankers to the US and soliciting business here.) Perhaps only UBS engaged in such conduct, and other banks only accepted US depositors but did not actively seek them out. We'll see.

    As tho the 100 depositors that Patrick is referring to, the number of depositors to whom letters have been sent likely originated from either CS or the Swiss government, both reliable sources.

  5. The Swiss newspaper Tribune de Geneve reported a few days ago that two adolescent children of a Geneva asset manager were kept for six hours and interrogated by CBP upon arriving in the US to visit their grandparents, regarding their father, his work, his visits to the US etc.
    The article also mentions that the names of around 10,000 employees of five banks had been handed to the Us with authorization of the Swiss gov't (but in violation of Swiss privacy laws) (the vast majority of these bank employees had no contact with US accounts.) A Swiss lawyer is now recommending to these employees that they not only not travel to the Us but also not outside Switzerland, as most countries have extradition treaties with the US. Many employees are now suing the banks for illegaly disclosing their identities. The five banks said to have delivered employee information are CS, Baer, HSBC, ZKB and BKB.
    Link: (in French)
    (Jack I am aware of rules for placing comments where relevant, did not know where else to place this.)


Please make sure that your comment is relevant to the blog entry. For those regular commenters on the blog who otherwise do not want to identify by name, readers would find it helpful if you would choose a unique anonymous indentifier other than just Anonymous. This will help readers identify other comments from a trusted source, so to speak.